Do blood tests cause anemia in hospitalized patients? The effect of diagnostic phlebotomy on hemoglobin and hematocrit levels.

OBJECTIVE To determine whether phlebotomy contributes to changes in hemoglobin and hematocrit levels in hospitalized general internal medicine patients. DESIGN Retrospective cohort study. SETTING General internal medicine inpatient service at a tertiary care hospital. PARTICIPANTS All adult patients discharged from the Toronto General Hospital's internal medicine service between January 1 and June 30, 2001. A total of 989 hospitalizations were reviewed and 404 hospitalizations were included in our analysis. MEASUREMENTS AND MAIN RESULTS Mean (SD) hemoglobin and hematocrit changes during hospitalization were 7.9 (12.6) g/L (P<.0001) and 2.1% (3.8%) (P<.0001), respectively. The mean (SD) volume of phlebotomy during hospital stay was 74.6 (52.1) mL. On univariate analysis, changes in hemoglobin and hematocrit were predicted by the volume of phlebotomy, length of hospital stay, admission hemoglobin/hematocrit value, age, Charlson comorbidity index, and admission intravascular volume status. The volume of phlebotomy remained a strong predictor of drop in hemoglobin and hematocrit after adjusting for other predictors using multivariate analysis (P<.0001). On average, every 100 mL of phlebotomy was associated with a decrease in hemoglobin and hematocrit of 7.0 g/L and 1.9%, respectively. CONCLUSIONS Phlebotomy is highly associated with changes in hemoglobin and hematocrit levels for patients admitted to an internal medicine service and can contribute to anemia. This anemia, in turn, may have significant consequences, especially for patients with cardiorespiratory diseases. Knowing the expected changes in hemoglobin and hematocrit due to diagnostic phlebotomy will help guide when to investigate anemia in hospitalized patients.